Privacy-Preserving GWAS Computation on Outsourced Data Encrypted under Multiple Keys Through Hybrid System

Abstract

Genome-wide association studies (GWAS) necessitate genomic information of a large population of an individual to achieve reliable results. The human genome can expose sensitive information and is potentially re-identifiable, which raises privacy and security concerns, making individuals deter from contributing their genomic information for such studies. Therefore, there is a need for secure and efficient analysis in which the data owners can securely allocate both the computation and storage on the untrusted cloud. Previous solutions used either a single key setting or far from being practical. Also, efficiency is a big limitation for real life applications. In this paper, we propose a novel hybrid solution that uses the concept of multi-key homomorphic encryption to encrypt ciphertexts using different public keys. To decrypt all the ciphertexts by using a single secret key, we add the re-encryption property of proxy re-encryption to multi-key homomorphic encryption. Also, our framework uses the recently introduced hardware-based architecture (i.e software guard extensions) to securely perform GWAS on genomic data in a privacy-preserving manner and ensure high efficiency by speeding up the computation while preserving the privacy of the data owners. To the best of our knowledge, our scheme is the first multi-key homomorphism combined with intel SGX for analyzing genomic data.